Function Generator Square wave / Pulse Isolator

[sourcecharge]

So, I am a recovering FG overvoltage victom that was caused by complete accedent, and could have been averted if I had paid more attention.

I am a firm believer in murphy's law, so I was sure that it was going to happen again.

Therefore, I took steps to solve the problem.

The problem of course is that I use the function generator in proto settings with leads that are hanging off of the bread boards, and also, the bread board voltage isolation between the pins are not always perfect.

The solution was of coarse to pay more attention to what I was doing but also to isolate the FG with a optocoupler, and then use a N and P channel mosfet to give only positive and/or negative output.

Which optocoupler did I use?

TLP352. 

It's the best one I could find that was through hole, but there is a better one that is surface mounted.

15-30Vcc output rails allows it to not only act as an isolator, but as an amplified output.  It does require a 100nF cap between the output rails which is annoying but not that big of a problem.  The mosfets are in their normal forward direction with the output of the optocoupler going directly to both mosfets' gates.

In order to maintain frequency responce, a 100 ohm resistance is used on the drain of each mosfet, but I'm using 3 300 ohm 1/2 W resisitors in parallel.

So, the five inputs are:
FG output and ground, which are isolated from the next three:
The negative, the positive and a ground.

The negative can be 0V, or the positive can be 0V as long as the Vcc total is 15V - 30V.

1k ohms in series with the input, and an extra 1n4150 diode in the forward position to protect the FG, the isolator works with down only 3 Vp positive square wave/pulse input all the way up to 10Vp positive square wave /pulse.

There are three outputs, 1 +, 1 -, and 1 with both a + & - swing.

I run it at +/- 10 Vp and have three outputs to choose from.

The bandwith with minimum distortion on this is about 0-5 MHz which I think is not bad for an optocoupler.

Anyways, just wanted to share the experience and to let others know about the solution that if you were going to have to buy the parts, it would be less than 5 bucks, compared to hundreds or even thousands of dollars it would cost to have get a new FG or fix the one you accedentally over voltaged like I did.

 

[t1d]

Good job. I think we all would be interested in pictures of the damage and repair, a rough diagram/schematic of the protection circuit you designed and pictures of that circuit assembly... More details about what led up to the event, too...

[sourcecharge]

Good job. I think we all would be interested in pictures of the damage and repair, a rough diagram/schematic of the protection circuit you designed and pictures of that circuit assembly... More details about what led up to the event, too...

Well, it's basically just an optocoupler, and it's not anything new or special.

Here is the link for the datasheet of the TLP352:

https://toshiba.semicon-storage.com/info/docget.jsp?did=12756&prodName=TLP352


On page 7 of the datasheet, you can clearly see exactly how to set it up. 

The only difference with what I did was I added 2 mosfets, 1 N channel, and 1 P channel. 

I use the rails at -10 and +10, while the mosfets switch between ground and their corresponding rail.

(N channel = +10, P channel = -10)

As for pictures, I really don't like the look of how I soldered it, and the box I used really was for something else completely.

On the outside, it's just a red plastic box, with 6 bannana plug sockets (three on each side), and 1 BNC connector for the FG.

The sockets and BNC connectors are labeled and it sits nicely next to my power supplies, and function generator.



The function generator repair was actually a long thread, but if you are interested in it here:

https://www.eevblog.com/forum/repair/sdg1025-overvoltage-and-relay-malfunction-repair/msg1841849/#msg1841849

If you want pictures of the repair, post in that thread of your interest.

[David Hess]

It pisses me off that newer test instruments like function generators do not have either a protected output, which is trivial to do once you know the trick, or at least a documented and repairable output stage.  It is one of the features I look for in the specifications and no test instrument newer than about 1990 qualifies.